Computational Intelligence - August 2016 - 53
synthesize the level of activation of each affective
dimension of a concept. Beyond emotion detection,
the Hourglass model is also used for polarity detection tasks. Polarity is defined in terms of the four
affective dimensions, according to the formula:
p = / Ni = 1
P (c i) + |At (c i)| - |S (c i)| + Ap (c i)
(36)
3N
Table 2 Error (in percentage) on the reference set exploiting different
losses and different MS strategies on AffectiveSpace 1.
ELMs
Loss
MS Method
BLB
SRC
SUS
BLHS
Pleasantness
L1
5.32 ! 0.16
5.95 ! 0.18
5.96 ! 0.19
4.76 ! 0.14
L2
5.85 ! 0.18
6.59 ! 0.21
6.57 ! 0.21
5.30 ! 0.17
L3
4.75 ! 0.14
5.21 ! 0.17
5.31 ! 0.16
4.16 ! 0.13
where P is the pleasantness, At the attention, S
L4
5.28 ! 0.16
5.92 ! 0.18
5.92 ! 0.19
4.75 ! 0.15
the sensitivity, Ap the aptitude, ci an input conL5
5.36 ! 0.17
5.88 ! 0.19
5.89 ! 0.19
4.77 ! 0.14
cept, N the total number of concepts, and 3 the
normalization factor (as the Hourglass dimensions are defined as f loats ! [- 1, 1]). In the equaTable 3 Error (in percentage) on the reference set exploiting different
tion, Attention is taken as absolute value since
losses and different MS strategies on AffectiveSpace 2.
both its positive and negative intensity values
ELMs
MS Method
correspond to positive polarity values (e.g., 'surLoss
BLB
SRC
SUS
BLHS
prise' is negative in the sense of lack of Attention,
Pleasantness
but positive from a polarity point of view). SimiL1
3.53 ! 0.11
3.93 ! 0.12
3.89 ! 0.12
3.14 ! 0.10
larly, Sensitivity is taken as negative absolute
value since both its positive and negative intensity
L2
3.82 ! 0.12
4.33 ! 0.14
4.32 ! 0.13
3.48 ! 0.10
values correspond to negative polarity values
L3
3.11 ! 0.10
3.46 ! 0.11
3.47 ! 0.11
2.74 ! 0.09
(e.g., 'anger' is positive in the sense of level of
L4
3.45 ! 0.11
3.90 ! 0.12
3.93 ! 0.13
3.13 ! 0.10
activation of Sensitivity, but negative in terms of
L5
3.54 ! 0.11
3.83 ! 0.12
3.92 ! 0.12
3.14 ! 0.09
polarity). The publicly available Sentic API (on
http://sentic.net/api) was used to obtain for each
concept the level of each affective dimension.
Table 4 Training time (in minutes) when different losses and different
According to the Hourglass model, the Sentic
MS strategies are exploited on AffectiveSpace 1.
API expresses the levels as numbers ! [- 1, 1],
ELMs
MS Method
which are eventually mapped into the associated
Loss
BLB
SRC
SUS
BLHS
polarity according to Eq. (36). In order to perform
a binary classification task for each affective dimenPleasantness
sion and polarity, the values are then discretized: +1
L1
15.08 ! 1.09
10.01 ! 0.76
10.04 ! 0.71
18.03 ! 1.27
for positive values and -1 for negative ones.
L2
15.10 ! 1.10
10.01 ! 0.77
10.07 ! 0.73
18.10 ! 1.22
The experiments eventually involve two tasks:
L3
15.04 ! 1.09
10.06 ! 0.77
18.10 ! 1.31
10.05 ! 0.70
❏❏ Classification of each affective dimension level
L4
15.07 ! 1.08
10.05 ! 0.73
10.05 ! 0.71
18.08 ! 1.20
and polarity detection for concepts expressed
L5
15
.
03
!
1
.
00
10.01
0.76
10
.
05
!
0
.
72
18
.11 ! 1.20
!
according to AffectiveSpace 1 [93];
❏❏ Classification of each affective dimension level
and polarity detection for concepts expressed
error on the reference set of the ELMs model selected by
according to AffectiveSpace 2 [94];
exploiting regularizer w 2 , different losses (L1, f, L5 in
In both cases, the dimension of the space M has been set
equal to 100.
Table 1), and different MS strategies (Bag of Little Bootstraps-BLB, Simplified Rademacher Complexity-SRC, Simplified Uniform Stability-SUS, and Bag of Little Hypothesis
IX. Experimental Results
Stabilities-BLHS). In Table 4, for AffectiveSpace 1, we have
In this section1, we show the results of applying the ELMs
reported the time required to build the ELMs model selected
models described in Section V to the Affective Analogical
by exploiting different losses and different MS strategies. In
Reasoning datasets described in Section VIII, where the perparticular, we reported only the time required for the Pleasformance of the models has been assessed by using the MS
antness task.
strategies described in Section VII.
From Tables 2, 3, and 4 we can state that:
In Tables 2 and 3 we have reported, respectively for AffectiveSpace 1 and AffectiveSpace 2 and for the Pleasantness, the
❏❏ AffectiveSpace 2 is able to better predict the affective
dimensions and polarity with respect to AffectiveSpace 1.
1
We do not report all the details and experiments because of space constraints, all the
❏
❏
BLHS is the best method to perform MS since it is the one
details can be found in the technical report available at http://sentic.net/slt-basedthat more often selects the most accurate model according
elm-for-big-social-data-analysis.pdf.
AUGUST 2016 | IEEE Computational intelligence magazine
53
http://www.sentic.net/api
http://www.sentic.net/slt-based
Table of Contents for the Digital Edition of Computational Intelligence - August 2016
Computational Intelligence - August 2016 - Cover1
Computational Intelligence - August 2016 - Cover2
Computational Intelligence - August 2016 - 1
Computational Intelligence - August 2016 - 2
Computational Intelligence - August 2016 - 3
Computational Intelligence - August 2016 - 4
Computational Intelligence - August 2016 - 5
Computational Intelligence - August 2016 - 6
Computational Intelligence - August 2016 - 7
Computational Intelligence - August 2016 - 8
Computational Intelligence - August 2016 - 9
Computational Intelligence - August 2016 - 10
Computational Intelligence - August 2016 - 11
Computational Intelligence - August 2016 - 12
Computational Intelligence - August 2016 - 13
Computational Intelligence - August 2016 - 14
Computational Intelligence - August 2016 - 15
Computational Intelligence - August 2016 - 16
Computational Intelligence - August 2016 - 17
Computational Intelligence - August 2016 - 18
Computational Intelligence - August 2016 - 19
Computational Intelligence - August 2016 - 20
Computational Intelligence - August 2016 - 21
Computational Intelligence - August 2016 - 22
Computational Intelligence - August 2016 - 23
Computational Intelligence - August 2016 - 24
Computational Intelligence - August 2016 - 25
Computational Intelligence - August 2016 - 26
Computational Intelligence - August 2016 - 27
Computational Intelligence - August 2016 - 28
Computational Intelligence - August 2016 - 29
Computational Intelligence - August 2016 - 30
Computational Intelligence - August 2016 - 31
Computational Intelligence - August 2016 - 32
Computational Intelligence - August 2016 - 33
Computational Intelligence - August 2016 - 34
Computational Intelligence - August 2016 - 35
Computational Intelligence - August 2016 - 36
Computational Intelligence - August 2016 - 37
Computational Intelligence - August 2016 - 38
Computational Intelligence - August 2016 - 39
Computational Intelligence - August 2016 - 40
Computational Intelligence - August 2016 - 41
Computational Intelligence - August 2016 - 42
Computational Intelligence - August 2016 - 43
Computational Intelligence - August 2016 - 44
Computational Intelligence - August 2016 - 45
Computational Intelligence - August 2016 - 46
Computational Intelligence - August 2016 - 47
Computational Intelligence - August 2016 - 48
Computational Intelligence - August 2016 - 49
Computational Intelligence - August 2016 - 50
Computational Intelligence - August 2016 - 51
Computational Intelligence - August 2016 - 52
Computational Intelligence - August 2016 - 53
Computational Intelligence - August 2016 - 54
Computational Intelligence - August 2016 - 55
Computational Intelligence - August 2016 - 56
Computational Intelligence - August 2016 - 57
Computational Intelligence - August 2016 - 58
Computational Intelligence - August 2016 - 59
Computational Intelligence - August 2016 - 60
Computational Intelligence - August 2016 - 61
Computational Intelligence - August 2016 - 62
Computational Intelligence - August 2016 - 63
Computational Intelligence - August 2016 - 64
Computational Intelligence - August 2016 - 65
Computational Intelligence - August 2016 - 66
Computational Intelligence - August 2016 - 67
Computational Intelligence - August 2016 - 68
Computational Intelligence - August 2016 - 69
Computational Intelligence - August 2016 - 70
Computational Intelligence - August 2016 - 71
Computational Intelligence - August 2016 - 72
Computational Intelligence - August 2016 - 73
Computational Intelligence - August 2016 - 74
Computational Intelligence - August 2016 - 75
Computational Intelligence - August 2016 - 76
Computational Intelligence - August 2016 - Cover3
Computational Intelligence - August 2016 - Cover4
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_202311
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_202308
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_202305
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_202302
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_202211
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_202208
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_202205
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_202202
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_202111
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_202108
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_202105
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_202102
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_202011
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_202008
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_202005
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_202002
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_201911
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_201908
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_201905
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_201902
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_201811
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_201808
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_201805
https://www.nxtbook.com/nxtbooks/ieee/computationalintelligence_201802
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_winter17
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_fall17
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_summer17
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_spring17
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_winter16
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_fall16
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_summer16
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_spring16
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_winter15
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_fall15
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_summer15
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_spring15
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_winter14
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_fall14
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_summer14
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_spring14
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_winter13
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_fall13
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_summer13
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_spring13
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_winter12
https://www.nxtbook.com/nxtbooks/ieee/computational_intelligence_fall12
https://www.nxtbookmedia.com